I have noticed that curve~'s curve parameter works the way it should from the range of 1 to -0.99999etc, but once it becomes -1 it actually curves the same way as if it were 1. Is this a bug?

curve-1-curves-same-as-1

it curves the same way, which is no curve at all... what did you expect?

Uh. One curve curves one way, the other is the inverse? Therefore one way it should curve up very fast, then slow down for the rest, the other way it should curve almost nothing at all until the end, THEN curve really fast? Again, it curves exactly how it should from -0.9999 through to +1.0, but if its -1.0 it all of a sudden flips around and does what +1 does?

Compare 0.5 to 0.999999 to 1.0, then compare -0.5 to -0.999999 to -1, then compare -1 to 1.

well that that happens when scientist write objects, they regulary forget the practice. it is probably explainable in regards of the math, but of course the input range should have been limited or mapped so that it does not occur.

otoh one might argue that you wil never use -1. (or 1. for that matter) because the curve will contain less than 1 sample which arent 0. or 1. - there is no curve anyway.

the third bug in this object is that it is not documented what it is doing inside.

i would like to know what -1. does, for example, then we could at least fix these bugs ourselves. it is not exp(a*log(90))-1)/(90-1) from what i can tell. the reference is talking of an "approximisation" which probably also does not mean it would use a lookup table.

the third bug in this object is that it is not documented what it is doing inside. 

i would like to know what -1. does, for example, then we could at least fix these bugs ourselves. it is not exp(a*log(90))-1)/(90-1) from what i can tell. the reference is talking of an "approximisation" which probably also does not mean it would use a lookup table.


curve~ -1 curves same as 1